Fuel cells for direct conversion of hydrogen or carbonaceous fuels to electricity have shown great theoretical promise, but have not become widely used in commerce because of technical problems and economic reasons. In the field of hydrogen-air/O.sub.2 fuel cells, power density, that is kilowatts of power generation per kilogram, has been marginal, and the lifetime has been unsatisfactorily short. Prior art cells have experienced drop-off in power with age due in part to poisoning of catalysts or electrolyte membranes, and the poor distribution of fuel gases internally has led to thermal hot spots leading to cell failure and the like.
A particularly important class of fuel cells with promise for stationary and mobile electricity generation is the low temperature H.sub.2 /O.sub.2 fuel cell employing solid polymeric proton exchange membrane having a noble metal catalyst coated on at both sides thereof, which membrane is located between the fuel cell electrodes. These fuel cells employ H.sub.2 as fuel, whether directly supplied as such or generated in association with the cell bychemical reaction, such as electrolysis, from metal hydrides or from reformed hydrocarbons. The oxidant is O.sub.2 or air where suitable. Water is required both for cooling and for humidification of the membrane, to keep it from drying out and becoming inefficient or structurally weakened through stretching and cracking. Typically, the anode side dries out first for a variety of reasons, including: electro-osmotic pumping from anode to cathode; supply of gases in excess of the electrochemical reaction rate; and the air or oxygen flow on the cathode side purges both the product water and the water vapor passing through the membrane from the hydrogen anode side. Accordingly, the fuel gases need to be humidified in the fuel cell stack to reduce the dehydration effect. The cooling water removes excess heat generated in the slow combustion of the catalyst-mediated electrochemical reaction in the cells, and is conducted external of the stack for heat exchange. In some designs the cooling water is used to humidify the reactant gases.
There are several suitable electrode membrane assemblies (EMAs) available for such low temperature fuel cells. One is from H-Power Corp of Belleview, New Jersey which employs a Pt catalyst coated on a polymer film, such as DuPont NAFION.RTM. brand perflourosulfonated hydrocarbon as the membrane. Alternatively, Dow Chemical provides a perflourosulfonated polymer which has been reported in U.S. Pat. No. 5,316,869 as permitting current densities on the order of 4000 amps/s.f. with cell voltages in excess of 0.5 V/cell, for a cell stack power density in excess of 2 kW/s.f.
A typical design of a currently available fuel cell stack is the Ballard Fuel Cell Stack of 35 active electrochemical cells, 19 thermal management cells, and 14 reactant humidification cells employing a Pt on NAFION-117 EMA in stacks of 1/4" thick graphite plates. The stack is reported to have an overall volume of 0.5 cu. ft. with a weight of 94 lbs and a 3 kW output from H.sub.2 and O.sub.2.
However, the graphite plates must be relatively thick to provide structural integrity and to prevent reactant crossover. That is, since the graphite is porous to H.sub.2 and O.sub.2, it must be at least 0.060" thick to reduce the permeation crossover to an acceptable level. Further, graphite plates are brittle. Thus, they are prone to crack as the cell stacks must be placed under compression to effect intra and inter-cell sealing to prevent reactant gas leakage. Graphite plates have low thermal and electrical conductivity which gives rise to hot spots and dead spots. They are also difficult to manufacture, especially the gas distribution channels. The stack output is relatively low, on the order of 0.03 kW/lb. In the example cited above, the number of inactive cooling and humidification cells almost equals the number of active electrochemical cells. This effectively doubles the number of gasketed seals required in a stack thereby decreasing stack reliability and performance.
The aforementioned U.S. Pat. No. 5,316,869 does not offer a solution to graphite plate cell stack design as it is concerned with microprocessor control of a closed loop system external to the stack.
Accordingly, there is a need for an improved fuel cell design, and methods of producing the fuel cells and operation thereof which overcome limiting problems of the prior art.